One of the assertions I added when solving #38809 fails with "do { return do { 1 } }", i.e. when the inner do block gets reaped at compile time. But there was no bug in this case to begin with, so we can just switch the assertion for a plain test. I added a regression test for this, but also some tests that checks that "do { return do { 1; @a }, do { 1; @b } }" returns as expected @a, @b in list context and the length of @b in scalar context. Patch against blead, tested ok. Vincent.Thread Next